Which type antenna should I use?

I need to design an antenna, but I am a newbie in antenna.
Could anyone give me some suggestion?
Requirements:
1. Freq band: 14GHz
2. Size: About Beer Can
3. Beam-width: not care,

Which type antenna should I use to get the maximum gain, except Lens?
many thanks.

Not yet enough information. Is the polarization circular? Is it for receive or transmit?

It has gain. It is a feed-horn. Many recipes are available on internet for simple pyramidal and circular feeds. If you connected a waveguide to a real beer can, you might even get some reasonable gain, but the main trick is to use suitable matching sections.

A fast way to get there is to find an existing feed, and scale it to the frequency you need. You don't even need expensive software.

Use the beer can as antenna, similar to how WiFi cantennas are designed. It is a simple construction and if you have a VNA available is it relative easy to find a good feeding point. Feeder is a lambda/4 monopole.
Find an old satellite dish head designed for 11-12 GHz and adjust it for 14 GHz is another possibility that probably can result in better gain then for the cantenna, maybe 12-15 dBi.
10 GHz is a popular amateur band with many antenna designs published, example: http://ra3wdk.qrz.ru/horn10368.jpg Scale all lengths to 14 GHz.
Helix antenna can give high gain, within a beer can size. Google for dimensions. Would probably be my choice if circular polarization not is a problem, or even an advantage.

Hi, All,

Thanks.
Circular polarization is OK.
Both for receiver and transmitter. TRX uses FDD, both in 14GHz band.
One clarification is the antenna size is the same as beer can, not means to use beer can as an antenna.
So Helix antenna has the highest gain, right?
I don't want to use horn, if do so, I have to use a dish.

Many thanks.

- - - Updated - - -

Hi,
I found two helix antenna calculators in internet, and the gain difference is about 20dB with the same input.
I don't know why, could you help me?

Input: 14GHz, 60 turns, turn sacing 0.1 lamda.
1. Link 1st: https://jcoppens.com/ant/helix/calc.en.php,get the gain 38.6dBi.
2. Link 2nd: https://www.daycounter.com/Calculators/Helical-Antenna-Design-Calculator.phtml, get the gain 18.6dBi.

Spacing 0.1 lambda between each turn seems too tight. I use 0.25 lambda.
38 dBi do not seem realistic, that must be something wrong. Helix with gain above 20 dBi requires very careful design. If higher gain then this is needed can an antenna array be designed.
Proposed satellite dish head with assumed gain +15dBi is without dish. It is a simple solution for reaching good performance. A Helix within similar external size will perform better, but there are more things that can go wrong, limiting peek performance.

Hi, Kafeman,

Thanks. I still have the following questions:
1. I want to use 0.2lamda, is it viable in 14GHz? that means 4.2mm spacing
2. If using 6GHz, could I use 0.08 lamda? that means 4mm spacing
3. Does the wave transmitted by helix antenna is circular polarization?

Best Regards,
Tony Liu

Not simple questions. Spacing for each turn affects antenna bandwidth, how pure the polarization will be and antenna impedance.
It is possible that reducing space between each turn can be used as a compensation for the dielectric in a plastic tube, if a such is used as a carrier for antenna structure.
These Helix calculators that can be found at Internet are general tools with emphasis on general. Recommend setting up a simulation for more correct answers.
Here can some additional info be found: https://www.antenna-theory.com/antennas/travelling/helix.php
If the antenna is designed for a well defined circular polarization, will it both receive and transmit with circular polarization.

E Kafeman said:

Spacing 0.1 lambda between each turn seems too tight. I use 0.25 lambda.
38 dBi do not seem realistic, that must be something wrong. Helix with gain above 20 dBi requires very careful design. If higher gain then this is needed can an antenna array be designed.
Proposed satellite dish head with assumed gain +15dBi is without dish. It is a simple solution for reaching good performance. A Helix within similar external size will perform better, but there are more things that can go wrong, limiting peek performance.

Click to expand...

Also for Tony.. 3 or 5 turns can illuminate easily a mis-applied 60cm TV satellite dish using offset feed without excessive spill-over.

Also possible is to use a long helix sans dish, but there is a wisdom about diminishing returns. Using 4 x 16dBi helix and a combiner network can easily work most satellites. It would look awesome, but hardly seems worth the trouble when a low-cost easy to put together simple small helix feed with a highly available dish product will perform better.

With all credit to Emerson, he has shown that many theoretical works (Kraus et. al), while being very good, do over-estimate the gain of helix antennas. Since the diagram is so widely known, I attach it here. It will allow folk to design any helix. To have a dimensionally non-critical, easy to make design, I advise to stay on the safe, robust side of the gain curve.

Do keep in mind that the impedance of a helix is said to be about 140 Ohms, though real structures above a ground plane disc might have it nearer 80 Ohms. Whatever it turns out to be, the need is for some kind of matching section. Some do it by having part of the first turn kept low over the ground plane. Some have a wide "tab" for the first quarter-turn or so, the tab supposedly being a "lecher-line" type of match, akin to microstrip, but with the dielectric being air. Some seem to have the helix suspended far from the groundplane with a straight conductor.

A quarter-turn is not the same thing as a quarter-wavelength, but if you get it right, it is easy to adjust. I favour having a quarter-wave matching section between the 50 Ohms connector and the beginning of the helix, it being placed behind the groundplane, like a fatter piece of tube with the connector on the end, but I have not had opportunity to try that idea (yet). Attached are pictures of the prototype.

Thank all of you.

- - - Updated - - -

If I use one right circular helix antenna and one left, I guess I can get about 2.5dB gain, right?

And how much space is needed between the two antenna, one lamda, or four/five lamda?

I just put two helix antennas separated by 39mm, one is left circular polarization, and the other is right circular polarization.
I hope polarization diversity can give me 2 dB gain, excluded the receiver side dual mode splitter insertion loss.

- - - Updated - - -

The working freq is 6GHz.
Is the mutual coupling between two helix is minimized by careful design?

tony_lth said:

I just put two helix antennas separated by 39mm, one is left circular polarization, and the other is right circular polarization.
I hope polarization diversity can give me 2 dB gain, excluded the receiver side dual mode splitter insertion loss.

The working freq is 6GHz.
Is the mutual coupling between two helix is minimized by careful design?

Click to expand...

OK - you folk have started me wondering some more about possible arrangements of helix.

1. The central support may be metallic without affecting the pattern if small enough diameter compared to the helix, but the helix may instead be supported on dielectric (plastic) tube which will affect the tuning some, or can have several kinds of dielectric centre diameters and shapes which will slow the wave and possibly generate a whole class of possible patterns.

2. For Tx / Rx where left-hand (LHCP) polarization is used in one helix, and right-hand (RHCP) is used in the other, and they are separated far enough to have insignificant mutual coupling, there might be 20 to 30dB of isolation (maybe? - I don't really know this yet). Probably the receive path would require additional transmit rejection using a multi-pole cavity bandpass filter, which might add 0.4 to 0.8dB of insertion loss.

3. Where two helix of the same polarization are used, spaced say 100mm (about 2*lambda for 6GHz), then a combiner between them could offer about 2.8dB additional gain, and the impedance of the combiner arms could be contrived to provide the impedance match at the same time if made so the lengths were some odd-number multiple of a electrical quarter-wavelength.

4. The helix(s) ..er.. is that "helices" might possibly be placed coaxially, with one slightly smaller than the other. Helix are wide-band enough that this might possibly work, and so obvious, someone might have tried it before, so there might be some existing literature.

5. If coaxial, then the first obvious try would be to make the helix angle 45° so that the currents in the LHCP helix are always at 90° to the currents in the RHCP helix at any plane distance from the start, but I am not sure this is necessary. We might be able to have all kinds of helix angle.

6. Circular feed-horns can support RHCP and LHCP simultaneously, and given that a helix is actually a traveling wave field-summing boundary structure, my imagination says "this is worth a try" and my ignorance says "It's a co-polar disaster waiting to happen and you should check this stuff from those who already know"!
Encouraging fields out of antennas is difficult enough without throwing into the same space, other fields on a different mission!

At 14 GHz, helical antenna results tiny and sensitive to mechanical fineties. Use a simple horn, conical or rectangular. Waveguide feed from a coaxial cable if needed. If gain is not needed, use only waveguide opening, some 90-degrees beam width.
I made such horns of hard paper an used aluminum kitchen foil for conductive surface, good results at no cost.

Thank all of you.
Another question, is it possible to transmit very good circular polarization wave by helix antenna?
such as axis ratio more than 20dB? Or even to 38dB.
I know in waveguide system it is possible to reach that performance with very good design.

I have not tried by myself but think a precision made helical an achieve close to 30 dB pol/copol ratio Another question is how to measure such quantity. You need a precision standard antenna and a fine test range, quite rare to get access to.